ontolog-forum
[Top] [All Lists]

Re: [ontolog-forum] So you want to be a Data Scientist?

To: ontolog-forum@xxxxxxxxxxxxxxxx
From: John F Sowa <sowa@xxxxxxxxxxx>
Date: Sat, 29 Dec 2012 09:41:54 -0500
Message-id: <50DF0132.4060003@xxxxxxxxxxx>
On 12/28/2012 9:22 PM, Hassan Aďt-Kaci wrote:
> This discipline of ours (CS) has been plagued with hyping for a long time.    (01)

That's what happens to academic disciplines that depend on grants
and contracts.  The academics publish peer-reviewed articles, which
can only be read and understood by other academics in the same field.
But when they compete for funding, they have to appeal to executives
who have no understanding of those publications.  Therefore, they
resort to meaningless slogans that sound good on the surface.    (02)

> To explain the nature of "Ontologies and the Semantic Web" in his
> contributed article (Dec. 2008), Ian Horrocks, a leading figure
> behind the theory and practice of Description Logics (DLs), employed
> analogous characters and language of the fictional Harry Potter
> children's novels.    (03)

I would dismiss that as silly if it weren't so misleading.
Their biggest buzzword is *decidability* -- but the real issue
is *computational complexity*.  They never mention that term,
because it would expose all their claims as nonsense.    (04)

> The trouble I see in such publications by influential members of
> the W3C is that one particular formalism—DL—is being confused with
> the general issue of formal representation and use of ontologies.    (05)

Not only that, OWL's major features are useless.  To prove that their
algorithms are decidable, they constrain all models to trees.  That
constraint rules out benzene rings (a huge part of organic chemistry
and pharmaceuticals).  It also rules out the metal or wood structures
of buildings, bridges, and airplane wings, which contain linked
triangles for rigidity.  And forget about electronic devices --
they all contain *circuits*.    (06)

But the size and computational complexity of trees grow
exponentially with depth.  Most programmers don't understand
decidability, but they certainly know that exponential algorithms
don't scale.  Decidability just means finite computing time --
even though that time may be longer than the age of the universe.    (07)

> Whether the various languages proposed by the W3C are able to fly
> beyond toy applications has yet to be proved, especially in light
> of the huge financial investment being poured into the semantic Web.    (08)

Ian and his DL friends claim that OWL is widely used.  But if you look
at 99% of the published OWL ontologies, you'll find that they don't go
beyond the subset of DL that Aristotle formalized with his syllogisms
2300 years ago.    (09)

But Aristotle's notation was much more readable than OWL.
He used controlled Greek, which later philosophers translated
to controlled Latin, controlled Arabic, controlled English,
and every other major NL.    (010)

Martin Hepp represented his GoodRelations ontology in OWL because
he could take advantage of the hype machine.  I give him a great
deal of credit for convincing Google and Microsoft to adopt it
for Schema.org.  But they translated it to a readable notation
that programmers (and logicians) could actually use.    (011)

> ... the goal is still to make semantic Web ontology languages work,
> no matter which method is used, as long as it is formal, effective,
> and efficient on real data.    (012)

I agree.  That's why I believe that Schema.org is an important step
toward liberating the Semantic Web from the clutches of OWL.  Their
ontologies are too weak to do much reasoning, but they aren't limited
by the useless constraints and cumbersome notation of OWL.    (013)

That means you can use Schema.org (including GoodRelations) with any
kind of logic -- or with legacy systems that have no formal logic
of any kind.    (014)

John    (015)

_________________________________________________________________
Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/  
Config Subscr: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/  
Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
Shared Files: http://ontolog.cim3.net/file/
Community Wiki: http://ontolog.cim3.net/wiki/ 
To join: http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J    (016)

<Prev in Thread] Current Thread [Next in Thread>